Perl 6 - the future is here, just unevenly distributed

IRC log for #openframeworks, 2015-11-01

| Channels | #openframeworks index | Today | | Search | Google Search | Plain-Text | summary

All times shown according to UTC.

Time Nick Message
00:25 bonsaielectric joined #openframeworks
00:47 i_n_g_o_ joined #openframeworks
00:49 bigx joined #openframeworks
01:35 w4ffles__ joined #openframeworks
01:48 i_n_g_o_ joined #openframeworks
02:49 i_n_g_o_ joined #openframeworks
03:22 vade joined #openframeworks
03:50 i_n_g_o_ joined #openframeworks
04:41 vade joined #openframeworks
04:51 i_n_g_o_ joined #openframeworks
05:51 i_n_g_o_ joined #openframeworks
06:04 meandi_2 joined #openframeworks
06:52 i_n_g_o_ joined #openframeworks
07:53 i_n_g_o_ joined #openframeworks
07:55 w4ffles__ joined #openframeworks
08:54 i_n_g_o_ joined #openframeworks
09:07 notjosh joined #openframeworks
09:55 i_n_g_o_ joined #openframeworks
10:55 i_n_g_o_ joined #openframeworks
11:19 halfdanj joined #openframeworks
11:19 halfdanj left #openframeworks
11:31 notjosh joined #openframeworks
11:41 i_n_g_o_ joined #openframeworks
11:52 kritzikratzi joined #openframeworks
12:07 zenx joined #openframeworks
12:07 zenx hi
12:27 bonsaielectric joined #openframeworks
13:30 bonsaielectric joined #openframeworks
13:59 bonsaielectric joined #openframeworks
14:26 bonsaielectric joined #openframeworks
15:13 vade joined #openframeworks
16:00 w4ffles__ joined #openframeworks
16:02 meandi joined #openframeworks
16:59 armadillu joined #openframeworks
17:06 kritzikratzi joined #openframeworks
17:07 Beliq joined #openframeworks
17:10 admsyn joined #openframeworks
17:22 kritzikratzi joined #openframeworks
17:32 tpltnt joined #openframeworks
18:01 ofzach joined #openframeworks
18:06 tpltnt hi ofzach
18:07 kylemcd joined #openframeworks
18:07 kylemcd hi everyone :) sorry i’m late!
18:07 Topic for #openframeworks is now Welcome :) Site: http://openframeworks.cc/ Logs: http://irclog.perlgeek.de/openframeworks/ Agenda: https://openframeworks.hackpad.com/IRC-Meetup-November-1-2015
18:07 tpltnt hi kylemcd
18:07 ofzach hey hey
18:07 roymacdonald joined #openframeworks
18:08 kylemcd lets wait a few more minutes in case arturo has a chance to join
18:09 kylemcd totally unreleated, i was just watching this https://www.youtube.com/watch?v=sB0xy74Zrj8
18:09 kylemcd i had no clue that realtime graphics looked like that in big games right now
18:10 roymacdonald joined #openframeworks
18:10 ofzach that is pretty nuts
18:10 ofzach I’m watching this final play from the duke miami game http://www.cbssports.com/collegefootball/eye-on-college-football/25359939/miami-pulls-off-miraculous-multi-lateral-return-td-to-win-vs-duke
18:10 admsyn definitely benefits from being kinda sorta on-rails :) fallout 4 is looking pretty plain last I saw
18:11 admsyn but still really impressive
18:11 ofzach I’m showing it to my mom right now — it’s a football play with 8 laterals…
18:12 ofzach I only play skate 3
18:12 kylemcd wow ok, now THAT is incredible
18:12 ofzach skate 3 is possibly the best game in the world.  my daughter and I like to ditch the skateboards and just walk around
18:13 admsyn :D
18:13 roymacdonald hi there
18:13 ofTheo joined #openframeworks
18:13 kylemcd hi roy :)
18:13 roymacdonald hey kyle, congrats for having your pull request merged :)
18:14 roymacdonald long time i’ve not been here at the IRC
18:14 kylemcd thanks :)
18:14 kylemcd well we have theo too, so let’s goo!
18:14 ofTheo :)
18:14 kylemcd the main thing is we just need to finish the change log
18:15 ofzach (I’m at a family event so I’ll be in and out of this IRC)
18:15 kylemcd i figured we could do that together right now
18:15 kylemcd thanks zach
18:15 kylemcd but before we go for it, is there anything else we need to talk about?
18:15 ofTheo RC3 :)
18:15 kylemcd oh, did something happen?
18:15 kylemcd i figured the next post would be the release
18:15 ofTheo I think we should do one more RC and if there are no issues then that becomes the release
18:16 ofTheo Thats usually how RC works
18:16 kylemcd great
18:16 kylemcd so we would post rc3 in the next couple days
18:16 ofTheo RC2 had a few issues on OS X and Windows - so its good to have a RC where there are no issues - the candidate becomes the release if it works well! :)
18:16 ofTheo yeah I think so - one video issue for OS X but mostly things look good
18:17 kylemcd then by the end of the week we announce that it’s the release, put up the blog post etc?
18:17 ofTheo yeah - I think that sounds good
18:17 ofTheo I would check with Arturo - but I think we're super close
18:18 kylemcd do you have a link for the osx issue you have in mind
18:18 ofTheo its almost finished
18:18 ofTheo https://github.com/openframeworks/openFrameworks/issues/4485
18:18 kylemcd ah right, i’ve been watching this
18:19 kylemcd it sounded like it was too hard to reproduce at first, then you found a way to trigger it
18:19 ofzach (is the android stuff that’s happening related to 0.9 ?  I think there are two things — x86 support and jonas was updating the code a bit…)
18:19 ofzach (sorry to interupt… the video stuff has been great to watch being solved)
18:19 roymacdonald joined #openframeworks
18:20 kylemcd i think jonas’ android work is up to arturo
18:20 kylemcd but it’s clear jonas tested it really well
18:20 ofzach cool I think these are good changes for android…
18:20 kylemcd whether it gets in 0.9.0 is on one more person to test it, it’s not holding us back from release
18:21 kylemcd besides the change log, the only other item on the agenda is moving to regular release cycle
18:21 kylemcd we have discussed this on and off, but i wanted to get feedback again on what the right cycle length would be
18:21 ofzach +1 to this, I think 6 weeks
18:22 ofzach it feels like a really small time frame is good
18:22 jvcleave I honestly prefer the big releases - seems like it gives time for platforms/IDEs to settle
18:23 micuat joined #openframeworks
18:23 kylemcd big releases could be fine too, but what time frame do you have in mind?
18:23 tpltnt why not combine both?
18:23 kritzikratzi from a user perspective, i don’t mind the irregular releases at all
18:23 kylemcd right now we’re operating on about an 18 month cycle
18:23 tpltnt I like the 6month release cycle of OpenBSD
18:23 ofTheo :) was just going to say
18:24 kylemcd kritzikratzi yeah it’s definitely more of something that would help with development rather than being meant for users
18:24 tpltnt it gives you enough time to plan ahead and work in changes
18:24 ofTheo I think minor releases could be automated though
18:24 tpltnt yes, agreed
18:24 ofzach I feel like minor changes are good and good if they are regular (ie, regular enough to put on a calendar)…
18:25 tpltnt i would opt for regular releases since a lot of artwork needs to be maintained
18:25 kylemcd in theory we could maintain the master for minor releases, and start running another branch in parallel for big features / changes
18:25 tpltnt and i feel oF has spread enough to consider it infrastructure
18:25 ofTheo @kylemcd I think thats the big challenge
18:26 ofTheo how to not break master when doing big / changes
18:26 kylemcd usually with big companies / products they ship everything, but don’t enable new features of it until they know there aren’t problems
18:26 ofTheo if we are pushing small stuff into another branch then we would need to pull that regularly into master / unstable
18:26 kylemcd unfortunately we don’t have the ability to “enable” things, the closest thing would be announcing them as “experimental"
18:27 tpltnt my company uses gitflow to handle releases etc. http://nvie.com/posts/a-successful-git-branching-model/
18:27 tpltnt so master is always a known good source
18:27 kylemcd tpltnt we started with gitflow when we moved to git ~4 years ago, but switched to our current model after a year or two
18:27 tpltnt ahh, ok
18:27 tpltnt why the change?
18:28 kylemcd part of the reason was that github didn’t support PRs to other branches at the time
18:28 tpltnt oh, i did not know that
18:28 kylemcd the bigger reason is that it was just enough overhead to get in the way
18:28 kylemcd i’m suggesting something inbetween our current model and gitflow, i think we could handle it now
18:29 kylemcd but, as a start, i think we should just set up a regular release schedule for master
18:29 kylemcd maybe between 2-4 months
18:29 tpltnt so you opt for 3?
18:29 kylemcd then when we have something we’re worried about releasing, we figure out the logistics then :)
18:31 kylemcd 2, 3, 4, something around there yeah
18:31 ofzach I would vote for 6 weeks or 8 weeks — I think if it’s really regular, you can plan when you can dedicate the time within those weeks for something you need to do…
18:32 ofTheo Would we do RCs for these? Or just hope there aren't regressions?
18:32 kylemcd any reason for 8 weeks vs 2 months? right now the irc meetings are on a monthly cycle and it seems to be the most succesful
18:32 kylemcd there would be no RCs
18:32 kylemcd instead we would have hotfix releases when there is a problem we didn’t catch
18:33 ofzach I just remember p5 used to do alot of releases, I think it will make the bigger ones easier and faster to do...
18:33 ofTheo So we would have 0.9.1 0.9.2 ... up to 0.9.11 etc - I guess if its every two months then a year would be 0.9.6 not too bad
18:33 ofTheo Yeah - I think a lot of the slow stuff is done
18:33 ofTheo lib updates might be something that could need coordination but otherwise
18:33 ofTheo should be good
18:34 admsyn +1 for short development cycle and hotfixes if necessary
18:35 kylemcd ok let’s make the next release (0.9.1) to be a fully automated release, and once it goes up we’ll kick off a 2 month cycle
18:35 ofzach yay
18:36 ofTheo sweet! :)
18:36 tpltnt kylemcd what automation is needed?
18:36 kylemcd the website doesn’t update automatically right now, i think
18:36 kylemcd some of the minor changes (like version number in the code) are still done manually
18:37 kylemcd one sec i can pull up the checklist that bilderbuchi prepared
18:37 kylemcd https://github.com/openframeworks/openFrameworks/wiki/Release-preparation-checklist
18:38 kylemcd parts of that process are automated (like the packaging has been automated for the nightlies)
18:38 kylemcd but not everything
18:39 kylemcd the regular release cycle will also make the milestone system irrelevant, unless we want to keep it for tracking bigger features or API changes
18:39 kylemcd so instead of having milestones for 0.9.1, 0.9.2 etc we would just have 0.10, 0.11, etc and the minor releases would be fully automated
18:40 jvcleave What I like about the big releases is that they seem to get a better quality of testing - something that I think OF needs
18:41 admsyn I think a counterpoint to that is that it makes it difficult for OF to keep up with the landscape shifting underneath it
18:41 admsyn i.e. whenever apple does anything ever
18:42 kylemcd yeah, 0.8.4 is severely broken on some platforms/configurations right now for example
18:42 kylemcd so even if it was well tested when it was released, we pay for it at the end of the cycle
18:42 kylemcd i think with the regular minor releases and irregular bigger version number changes we can get the best of both worlds
18:42 admsyn what if we do an RC and heavier testing on every other release on the shorter cycle?
18:42 jvcleave I guess I am not seeing how a timed cycle would help against that
18:43 admsyn gstreamer does something like that, and I think it works well
18:43 jvcleave I would think we would get half-baked releases that only covered some of the issues
18:43 ofTheo I think small releases that are automated HAVE to be stable and small bug fixes only
18:44 ofTheo bigger releases need a testing / RC period
18:44 kylemcd ideally, the automated release system is so smooth that there is also incentive to push things to master
18:44 jvcleave yeah - maybe it is just a clarification of what the goal of smaller/faster releases are
18:44 kylemcd otherwise we will end up in the same situation, with a separate branch containing all the really important up-to-date stuff
18:45 ofTheo the only other issue with automated releases is it puts a lot more burden on checking the quality of PRs
18:45 ofTheo ie: making sure a fix is tested on all platforms before merging
18:45 kylemcd i think with the automated testing we’re closer to making that happen, theo
18:45 ofTheo but I guess we should be doing that anyway :)
18:46 Banana_ joined #openframeworks
18:46 ofzach I feel like we should build a test period into the 2 month period, where we really push people to test…  (maybe by week 6 out of 8….)
18:46 ofTheo well - we can always try this new approach and if it doesn't work - look at other methods
18:46 Banana_ Hi there!
18:46 ofTheo ofzach - I agree
18:46 ofzach and in the test period really ping people to look at nightlies, etc
18:46 ofTheo it would be great if we could lock PRs somehow at that point
18:46 kylemcd hi banana_
18:47 kylemcd it’s nice to imagine people sitting down and testing OF, but in my experience that’s not what happens
18:47 Banana_ I've got a little question, i'd like to build a painting application, but with "dancing" strokes :)
18:47 kylemcd what actually happens is that there are a ton of people who use master, and pull regularly, and when they find something wrong they report it
18:47 Banana_ together with all the variables from the wacom (tilt, pressure, ..)
18:47 Banana_ Anyone has a suggestion as to where to start researching?
18:47 kylemcd then when there is an RC, one person (sometimes) goes through it carefully on one platform
18:47 ofTheo People do test OF RCs :)
18:47 jvcleave i think it does happen with RCs
18:48 ofTheo Nick and I have done it several times now and I see  that others do it when we post RCs to the forum
18:48 kylemcd banana_ maybe check https://github.com/mattebb/ofxTablet
18:48 ofTheo The main thing is to make it easier to test
18:48 ofzach @Banana_ we are in the middle of a conference right now :)
18:48 kylemcd and more generally, check http://www.ofxaddons.com/ :)
18:48 ofTheo On windows you can build all examples with a script but not run all examples
18:48 Banana_ Thanks! An i'll let you do your thing :)
18:48 ofTheo so this makes it slower to test
18:49 kylemcd in my opinion, it’s a counterproductive approach to expect people to test OF (except for pulling regularly and reporting issues)
18:49 kylemcd or at best, it’s an interim solution
18:49 ofTheo I totally disagree
18:49 kylemcd aif it happens at all, it should be the final step of a more complete automated testing platform
18:49 ofTheo We have found so many bugs before release by testings RCs
18:50 ofTheo me personally - but also the call to have people try RCs has been super helpful
18:50 ofTheo especially on platforms that aren't used as much by people on the dev list
18:51 ofTheo Its quality control - and nothing looks worse than a shiny new release that doesn't work
18:51 ofTheo its really the least we can do :)
18:51 kylemcd what kind of testing do you have in mind that shouldn’t be automated?
18:51 jvcleave ultimately it would be nice to have team/section leaders “sign off” on things
18:51 ofTheo I'm taking about the current process of RC releases.
18:51 jvcleave yeah - OF is game-like in that way - needs eyes - not just to compile
18:52 ofTheo and using the runAllExamples.sh scripts to quickly look at all the examples
18:52 ofzach I think having coverage is good, like how nick notices all the 10.8 issues….
18:52 tpltnt unit testing could be done, but clicking on things is harder
18:53 ofzach that’s the kind of stuff automated testing can’t find
18:53 ofTheo was thinking of something really dumb - but taking screen shots of all the examples and comparing them to the screenshots of a previous release
18:53 ofTheo pixel differences would indicate something to look at -
18:54 admsyn does every release have to count as a shiny new release though? I feel like we're discussing the difference between a x.x.x release and a x.x release
18:54 ofTheo wouldn't work for videoGrabber examples but maybe 75% :)
18:54 ofTheo I'm talking for major releases
18:54 kritzikratzi also doesn’t work for audio
18:54 ofTheo I think for minor / automated ones we need to make sure we're more careful not to introduce regressions as we won't be testing as much
18:55 ofTheo I think it might be hard to get people to really Test releases every two months - so we need to factor that in
18:56 ofTheo But we could start expanding the unit tests and maybe the screenshot idea is not actually that terrible of way to catch some issues? :)
18:56 admsyn I think auto-screenshoting would help a lot
18:57 kylemcd yes, i definitely can’t imagine thorough manual testing on a 2 month cycle.
18:58 ofTheo anyway I am ++1 trying the automated schedule - I think we will learn a lot and it might benefit OF in unexpected ways.
18:58 ofTheo we just need to set it up right - so that might need to be a slightly more in depth discussion maybe on the dev list?
18:58 kylemcd yeah, i’d be happy to start that up again
18:59 kylemcd i’m trying to find an example of another toolkit that does a regular release cycle but also asks for regular manual testing
18:59 kylemcd does anybody know one?
19:00 jvcleave i think you have to look at something like games
19:00 tpltnt OpenBSD asks for manual testing
19:00 ofTheo I think its definitely a good reason to work on automated tests.  finding clever ways to catch issues that aren't just compiler / linker
19:00 ofTheo the travis stuff is really helpful for that
19:01 ofzach (sorry I have to run!  I’m hanging out with fam right now…. ciao everyone  !!)
19:01 kylemcd bye zach
19:01 tpltnt bye zah
19:01 ofzach left #openframeworks
19:01 kylemcd i’d be curious to see how OpenBSD makes it work
19:01 kritzikratzi blender does manual testing: http://wiki.blender.org/index.php/Dev:Source/Architecture/Automated_Testing
19:01 kylemcd nice, blender is a great example
19:02 jvcleave yes - nice
19:02 jvcleave they also probably deal with a lot of examples
19:03 kylemcd ok i’ll send that email to the list and we can continue there
19:04 kylemcd but for now… :)
19:04 kylemcd this is the PR list https://gist.github.com/kylemcdonald/247bbecc8c0fae0af869
19:04 kylemcd i’m thinking it will be easiest to make one shared google doc
19:04 kylemcd and then we can paste the changes into github
19:04 kylemcd rather than having a bunch of PRs for the changelog
19:04 kylemcd so let me make that quickly
19:05 ofTheo One other thing that could be helpful going forward is to maybe do a release where we prune the examples
19:06 ofTheo There are a few quite legacy examples
19:06 admsyn +1
19:06 tpltnt +1
19:06 kylemcd good idea. maybe open issue for pruning the examples?
19:06 ofTheo and we could make a few examples that fill in the gaps
19:06 ofTheo will do :)
19:07 kylemcd https://docs.google.com/document/d/123bN3_8xE0gs7XLOs0-6NoNg42XvyClinh4m2X1Lz5k/edit?usp=sharing
19:07 jvcleave I’ve had it shot down before but I’ll throw out the idea again :) - break out the examples into a seperate repo that gets brought in during packaging
19:08 admsyn if that'd make it possible to just git clone the examples w/ project files than +many
19:08 ofTheo I'm allergic to submodules though :D
19:09 jvcleave i think git is allergic to them as well but the packaging process is scripted
19:10 jvcleave that way OF master is just the library
19:10 kylemcd jvcleave that could be good, i think anything we can do to separate release cycles and PR management of different parts of OF is good
19:10 ofTheo I would like to find a way to make it people who download OF via github to get the project files
19:10 jvcleave I also think it allows others to more easily contribute examples
19:10 jvcleave without having to but a PR against master
19:11 ofTheo - I can imagine that a lot of people don't know how to get the examples generated.
19:11 ofTheo might be good to just point a link to the new PG app which seems to work really well
19:11 kylemcd i would like to see project files completely disappear from all OF repos :)
19:12 kylemcd drag and drop folder to PG + auto open in IDE = no more project files
19:13 admsyn that'd be great
19:13 ofTheo I would like to see submodules disappear too ;)
19:14 armadillu joined #openframeworks
19:15 kylemcd something that wouldbe really useful
19:15 kylemcd i have some aliases in my bash profile for git
19:15 kylemcd like:
19:15 kylemcd alias push='git push origin HEAD'
19:16 kylemcd which pushes whatever branch you’re on right now
19:16 kylemcd alias ga='git add -u'
19:16 kylemcd alias gc='git commit -m'
19:16 kylemcd etc
19:16 jvcleave alias yolom="git push origin master"
19:16 ofTheo :)
19:16 admsyn your yolo needs a force
19:16 jvcleave good point
19:17 ofTheo git add * && git commit -a :)
19:17 kylemcd :) i’ve found that HEAD is better than master because you can be working on a different branch and you don’t have to specify it
19:17 kylemcd scary
19:17 kylemcd what would be great is a one-line alias for “push all my changes, and all the submodule changes"
19:17 ofTheo those are useful though!
19:17 kylemcd and another one for “pull all changes, and submodule changes"
19:17 kylemcd since that’s what you want to do with submodules most of the time
19:17 bonsaielectric joined #openframeworks
19:18 kylemcd i think: alias pull=‘git pull origin HEAD && git submodule sync && git submodule update’ ?
19:18 kylemcd or something similar
19:18 ofTheo yeah - I think stuff like this + scripts for grabbing the latest PG could be really good to have on hand
19:18 kylemcd someone with more git-fu would know how to handle the submodule init part...
19:21 nervous joined #openframeworks
19:21 admsyn is there a way to check the boxes in that gist kylemcd ?
19:21 admsyn or otherwise say " this PR is accounted for in the changelog "
19:21 kylemcd oh noooo
19:21 kylemcd only i can check them
19:21 kylemcd one sec
19:22 tpltnt kylemcd is the idea about multiple blog posts about 0.9.0 still on?
19:22 tpltnt i do have more time right now
19:22 kylemcd yes! can you help with that?
19:22 tpltnt sure
19:22 kylemcd admsyn https://github.com/openframeworks/openFrameworks/issues/4104#issuecomment-151997326
19:23 admsyn thx
19:24 admsyn whenever I check a box it makes the rest unclickable github pls
19:25 kylemcd it looks like it works if you reload the page
19:25 kylemcd but yeah that’s rough
19:27 tpltnt adamsyn: i pasted all PRs into the doc
19:27 tpltnt so they can be removed when addressed
19:27 tpltnt ugly, but should work
19:29 kylemcd great. better solution tpltnt. thanks!
19:30 kylemcd if you’ve finished one, remove it. i’ll remove the ones that were checked right now.
19:30 bilderbuchi joined #openframeworks
19:32 tpltnt you're welcome
19:32 ofTheo hey @bilderbuchi!
19:32 bilderbuchi hey
19:33 kylemcd hey bilderbuchi :)
19:33 kylemcd we’re going through the PRs and changelog now https://docs.google.com/document/d/123bN3_8xE0gs7XLOs0-6NoNg42XvyClinh4m2X1Lz5k/edit?usp=sharing
19:33 bilderbuchi hey. kinda late to the party, I guess :D
19:33 bilderbuchi yeah I just saw. man this list is huge (knew that before, but still...)
19:34 notjosh joined #openframeworks
19:35 bilderbuchi I got some other things on my plate right now, so don't know yet how much I can contribute, but I thought I'd drop by if you need anything specific from me.
19:36 bonsaielectric joined #openframeworks
19:36 kylemcd i think we’re good. the plan is to finish this changelog, release RC3, and promote it to 0.9.0 after a few days of no bugs.
19:37 kylemcd tpltnt is preparing the ofSite tutorials for the 0.9.0 blog post
19:42 bilderbuchi sounds good
19:46 admsyn alright, done my changelogs \o/
19:47 bilderbuchi *highfive*!
19:48 kylemcd :)
19:49 notjosh joined #openframeworks
19:49 kylemcd admsyn this was the PR for the big sound change https://github.com/openframeworks/openFrameworks/pull/3479
19:49 kylemcd it looks like you’re adding info about that
19:49 admsyn yeah just caught that :)
19:49 admsyn I was like "I'm sure there's more than that though.."
19:51 admsyn didn't there used to be an "examples" section on the previous changelogs?
19:52 admsyn nope, guess it can go under the relevant section though
19:52 ofTheo joined #openframeworks
19:55 bilderbuchi skimmed the irc log, a couple of remarks:
19:55 bilderbuchi 1) I think full release automation is essential. The only thing you should have to do is change the version number in ofConstants, create a tag, and push/merge to master, the rest is on rails. this should really really help with many regular releases (but there seems to be consensus on that anyway)
19:56 bilderbuchi 2) the new milestones, in a timed-release model, could be, instead of numbers, next-major, next-minor, next-patch, so that you pick stuff from that list, an when release time comes you check what kind of issues got closed since last (which would also automatically determine the next version number (i.e. a major/minor/patch bump). also, all the rest of the issues in the respective milestones just carries over to the _n
19:56 bilderbuchi ext_ release. less pressure, less maintenance that way.
19:57 kylemcd bilderbuchi i agree with both points
19:57 admsyn in the changelog, there should be something documenting the general shift from getIsWhatever() to isWhatever() I think?
19:58 admsyn I don't remember where we landed on that, but there's a few PRs I can spot that are like "changed deprecated getIsX to isX"
19:58 kylemcd we could have a note in the “general” section at the top
19:58 kylemcd but i think it’s still important to document it for each case where it comes up
19:59 kylemcd also: does everyone agree that there are some PRs like “fixed typo” where they don’t need to appear in the changelog?
19:59 kylemcd if so, i can remove some of these
19:59 admsyn yes I've been skipping over / removing those
19:59 kylemcd great
20:01 banana_ joined #openframeworks
20:02 notjosh joined #openframeworks
20:02 bilderbuchi kylemcd the reason I didn't put checkboxes in the PR list gist that I uploaded was that I knew only I could check them. I see you also found this out in the meantime ;-)
20:04 kylemcd ;) yep
20:05 admsyn alright, addressed all the remaining sound ones I believe
20:05 admsyn I'm off, happy 0.9 everyone!
20:07 banana_ hmm, time for the noob to annoy you all!
20:07 banana_ :D
20:07 bilderbuchi regarding the "screenshot-based" testing you guys talked about earlier: Be aware that on Travis you can enable an xvfb (a virtual framebuffer), as far as I understood it, you can run graphical output with that you could compare/analyze. I think people even do browser output testing that way, so could be something we should look into
20:07 bilderbuchi http://docs.travis-ci.com/user/gui-and-headless-browsers/#Using-xvfb-to-Run-Tests-That-Require-a-GUI
20:11 ofTheo nice @bilderbuchi - building it in to what we already have would be ideal
20:12 bilderbuchi yeah. I never tried it, and our usecase could be just too complex, but it seems to be worth investigating.
20:13 bilderbuchi possible problem could be if stuff needs a dedicated GPU or some other hardware that's just not available on the server boxes travis runs on
20:14 kylemcd hey banana_ what’s up? :)
20:18 banana_ Firstly, i don't if what i want to accomplish is way to difficult for someone with no c++ experience :)
20:18 banana_ but here it goes:
20:19 banana_ would it be possible to create a kind of series of generative visuals, with the variables being input from the wacom (will be a constant after a mark) and from the music (always changing), and then combine their output in "viewport" with photoshop alike layer control?
20:20 banana_ Basically i want something like this: https://www.youtube.com/watch?v=ALlaQWVXcEg
20:20 banana_ but then instead of using dozens of sprites, more complex algorithms
20:23 kylemcd it’s definitely possible
20:23 kylemcd but are you somewhere that you can find other people using c++ creatively?
20:23 kylemcd otherwise i would recommend using another toolkit for this kind of thing
20:23 banana_ not really
20:24 banana_ the only "creative developers" i know are making sites with wordpress... ;)
20:25 banana_ and, i've been looking for such toolkits but find it hard to find one..
20:25 kylemcd this would be a good place to use vvvv or processing i think
20:26 kylemcd or maybe even nodebox
20:26 banana_ i do have a lot of experience in grasshopper for rhino
20:26 kylemcd vvvv might feel familiar then
20:26 banana_ is essentially a lot like node box or vvvv
20:27 banana_ but performance might be an issue i think
20:27 kylemcd yeah grasshopper wouldn’t work for this :)
20:27 kylemcd but it does mean you’re on windows, right?
20:27 kylemcd so vvvv would be my first suggestion
20:28 banana_ Not on windows, but i did use it alot last year
20:28 banana_ vvvv might work
20:28 kylemcd that said, you can definitely do it with OF — i just suggest finding someone who you can ask c++ questions
20:29 kylemcd OF makes this kind of thing much easier than it would be otherwise, but it’s still not like using photoshop or even grasshopper
20:29 banana_ yeah
20:30 banana_ the good thing would be that i can really control the output in every way, the bad thing is that i can really control the output in every way
20:30 banana_ :P
20:30 kylemcd i think to replicate the video you linked to, you could probably write less than 300-400 lines of code
20:30 kylemcd and you wouldn’t have to worry as much about performance limitations compared to other toolkits
20:30 banana_ + a sprites library?
20:30 kylemcd you would just use ofImage
20:30 banana_ i can try that for starters
20:31 banana_ it's basically just layering stuff on top of stuff and sometimes flattening things to free up resources, right?
20:31 kylemcd yes
20:32 kylemcd the blending / flattening is the thing that’s harder with OF than photoshop
20:33 kylemcd there are some default blend modes with ofEnableBlendMode
20:33 banana_ yeah, would this be cpu intensive or can openGL generate that for me?
20:34 kylemcd it would all be with opengl behind the scenes
20:34 banana_ it's gonna be a tough project either way :P
20:34 kylemcd you can use an ofFbo as your “canvas” if you want to render off-screen, or into a “flattened” layer you blend somehow
20:34 kylemcd look into https://github.com/patriciogonzalezvivo/ofxFX for other effects
20:35 banana_ i will, thanks!
20:35 banana_ would this also be realistic in processing?
20:35 banana_ performance-wise etc
20:37 banana_ & what GUI would you suggest?
20:38 kylemcd it’s realistic if you’re using opengl mode, and there is a wacom library (there probably is)
20:38 kylemcd for OF, ofxGui is fine for 90% of things
20:43 banana_ ofxGUI looks good
20:43 banana_ does openFrameworks make c++ feel like scripting?
20:43 banana_ i'm more a scripter than a programmer :)
20:44 kylemcd how would you describe the difference in feeling between scripting and programming?
20:46 banana_ lots of standard functions that you can combine to do different things
20:46 banana_ programming would be more like defining what it is, where to put it in the memory, etc etc, very mathematical
20:50 kylemcd i see. i would say with OF you can use a lot of standard functions to do different things related mostly to making visuals.
20:50 kylemcd but sometimes to do something new or different you need to have more precise control, which you can get if you need it
20:50 kylemcd i would look through the examples to get a better feeling for it
20:51 banana_ i just found a 40-videos series on youtube :)
20:51 banana_ i'll go through then
20:51 banana_ them
20:52 banana_ & thanks for the info so far!
20:52 kylemcd great! good luck :)
20:53 banana_ Thanks :)
20:56 bilderbuchi ok got to go now. see ya guys!
20:58 kylemcd see ya bilderbuchi!
21:00 meandi joined #openframeworks
21:28 kylemcd omg there are so many PRs
21:28 Topic for #openframeworks is now Welcome :) Site: http://openframeworks.cc/ Logs: http://irclog.perlgeek.de/openframeworks/
21:37 banana_ joined #openframeworks
21:38 bilderbuchi left #openframeworks
22:13 kylemcd signing off for a bit, coming back to this later..
23:14 roymacdonald joined #openframeworks

| Channels | #openframeworks index | Today | | Search | Google Search | Plain-Text | summary